Improve style

demos/PyTorchNiftyNet/libs/dataset_niftynet.py

@@ -1,15 +1,17 @@
 import torch
 from torch.utils.data import Dataset
 
+
 class DatasetNiftySampler(Dataset):
     def __init__(self, sampler):
         super(DatasetNiftySampler, self).__init__()
         self.sampler = sampler
 
     def __getitem__(self, index):
         data = self.sampler(idx=index)
-        return torch.from_numpy(data['image'][..., 0, :]).float(),\
-               torch.from_numpy(data['label'][..., 0, :]).float()
+        image = torch.from_numpy(data['image'][..., 0, :]).float()
+        label = torch.from_numpy(data['label'][..., 0, :]).float()
+        return image, label
 
     def __len__(self):
         return len(self.sampler.reader.output_list)

demos/PyTorchNiftyNet/libs/loss.py

@@ -10,28 +10,18 @@ def __init__(self):
     def forward(self, output, label):
         probs = output.view(-1)
         mask = label.view(-1)
-        #eps = 0.00000001
         smooth = 1
-
         intersection = torch.sum(probs * mask)
-
         den1 = torch.sum(probs)
         den2 = torch.sum(mask)
-
-        #soft_dice = ((2 * intersection) / (den1 + den2 + eps))
-
-        soft_dice = ((2 * intersection) + smooth) / (den1 + den2 + smooth)
-
+        soft_dice = (2 * intersection + smooth) / (den1 + den2 + smooth)
         return -soft_dice
 
 
-def dice(input, target):
-
-    eps = 0.00000001
 
+def dice(input, target):
+    epsilon = 1e-8
     iflat = input.view(-1)
     tflat = target.view(-1)
     intersection = (iflat * tflat).sum()
-
-    return ((2. * intersection) /
-                (iflat.sum() + tflat.sum() + eps))
+    return 2 * intersection / (iflat.sum() + tflat.sum() + epsilon)